Biliary FACs as Risk Factors for Hepatic and Renal Lesions in Individual Fish
The following section presents the results of logistic regression analyses relating liver and kidney lesion occurrence in individuals of the primary target species from which bile was collected, as matched with levels of FACs-L (naphthalene wavelengths) and FACs-H (benzo[a]pyrene wavelengths) in bile of the same individuals, while simultaneously controlling for the biological risk factors of age or estimated age and gender. The critical level of significance for increased risk of lesion occurrence due to biliary FAC levels was set at p < 0.05. The reader is reminded that FAC levels estimate concentrations of metabolites of fluorescent aromatic compounds in bile (Krahn et al. 1986a) and are reflective only of recent exposure to polycyclic aromatic compounds (Collier and Varanasi 1991); specifically, levels of FACs-L estimate exposure to and metabolites of AHs with 2 to 3 aromatic rings, while FACs-H estimate levels of metabolites of AHs with 4 to 6 aromatic rings. Furthermore, because bile was collected and analyzed only from approximately 30% of the individuals examined for histopathology, this analysis is significantly limited in that it reflects only a portion of the histopathology data available from individual fish. This analysis was done because it was the only opportunity available to examine the influence of contaminant (AHs) exposure on the risk of lesion occurrence in individual fish.
Flathead sole--As applied to the 100 flathead sole for which bile was collected, this analysis failed to show any significantly increased risk in any hepatic lesion category attributable to levels of either category of biliary FACs (data not shown). However, an increased risk of necrotic lesions in the kidney was attributed to levels of biliary FACs-H (p = 0.038, grand mean = 0.193E-1); risk of renal necrosis increased by 1.928 for each 100 additional units of value (ppb, wet weight) for FACs-H.
English sole--Analyses of liver lesion occurrence in 166 individual English sole showed an elevated risk attributable to levels of biliary FACs-L (Table 13) only for the category of specific degenerative/necrotic lesions (SDN). This lesion was most frequently encountered in sole from sites in the urban embayments in Puget Sound--Elliott Bay and Commencement Bay.
Neither of the two indices of AH exposure (FACs-L and FACs-H in bile) were associated with an increased disease risk for any kidney lesion category.
Starry flounder--Neither measure of biliary FACs were risk factors for any hepatic lesion in the 274 starry flounder from which bile was collected. An increased relative risk of kidney necrosis was shown for biliary FACs-L (p = 0.05, grand mean = 0.2116E-1); risk of this lesion increased by 1.100 for each additional 1,000 units of value (ppb, wet weight) .
Hornyhead turbot--Analysis of hepatic and renal disease risk as related to biliary FACs levels in 112 hornyhead turbot in the data set failed to show increased risk for any lesion category in either organ.
White croaker--Analyses in white croaker demonstrated increased risks for both foci of cellular alteration and SDN as associated with levels of biliary FACs-L (Table 14). No increased risk for any lesion type was associated with biliary FACs-H . Neither biliary FAC measure was a risk factor for any kidney lesion in this subset of individual fish (n = 375).
Black croaker--No increased risk of hepatic or renal disease could be attributed to levels of biliary FACs in the small sample of 37 black croaker from which bile was collected.
Assessment of Variables Related to Contaminant Exposure (Determined on a By-Site Basis) as Risk Factors for Hepatic Lesions
The following section presents the results, by species, of multivariate logistic regression analyses testing the significance (p < 0.05) of various xenobiotic chemical exposure parameters (all variables shown in Table 3), determined for each year's sampling occurrence at the sites, in sediments (potential exposure), liver tissue (hepatic bioaccumulation), stomach contents (dietary uptake), and bile (estimate of recent exposure to and metabolism of AHs), in accounting for the variability in site- and year-specific hepatic lesion prevalences, while simultaneously controlling for the biological risk factors of mean age and gender ratio. The section for hepatic lesions is followed by a presentation of the results of identical analyses for kidney lesions in the same species.
Whenever both the low molecular weight aromatic hydrocarbons (LAHs) and high molecular weight aromatic hydrocarbons (HAHs) in either the sediments or stomach contents are identified as risk factors of hepatic or renal lesions, the total aromatic hydrocarbons (TAHs, representing the sum of LAHs + HAHs) are not mentioned in the text as risk factors. However, in these cases the TAHs are shown as risk factors in the tables.
Flathead sole--No risk factors among the chemical exposure parameters determined at each site of capture and year of sampling over Cycles I-V were identified for any detected hepatic lesion category (data not shown). Measures of chemicals in sediment and the biological variables were available from 13 sites over 1984-88; sample sizes for measures in liver tissue, stomach contents, and biliary FACs at the sampling sites were 10, 7, and 11, respectively.
English sole--Similar analyses in English sole showed a large number of chemical exposure-related risk factors for most of the hepatic lesion categories detected in this species (Tables 15-17).
Sediment chemistry--Of the chemicals measured in sediment (Table 15), the LAHs, HAHs, polychlorinated biphenyls (PCBs) and Metals 1 were highly significant risk factors for all detected liver lesion categories, with the exception of necrosis, which showed no risk factors. As individual risk factors, these variables accounted for between 38 and 68% of the intersite variation in prevalence among these hepatic lesions. Total DDTs and chlordanes were risk factors for SDN and proliferative lesions, but explained less of the variation (12-18%) than the above factors. Dieldrin was also a risk factor for both neoplasms (35%) and proliferative lesions (20%). It is important to note here that many of these sediment chemistry measurements, especially the LAHs, HAHs, PCBs and Metals 1, were highly and significantly intercorrelated (Table 18).
Stomach contents chemistry--Similar risk factors of hepatic disease in English sole were also identified in the chemicals measured in stomach contents as indicators of dietary exposure (Table 16). LAHs, HAHs, DDTs and PCBs were risk factors for prevalences of foci of cellular alteration, SDN, and proliferative lesions, but not for neoplasms. The only risk factor for hepatic neoplasms was chlordanes, which also significantly accounted for the variation in prevalence of SDN and proliferative lesions. Metals 1 was a risk factor for foci of cellular alteration, SDN, and proliferative lesions and explained a substantial proportion of the intersite variation in prevalence in these lesion types (72-78%). The chemicals measured in stomach contents that were risk factors for hepatic disease were also significantly intercorrelated, with the exceptions of the AHs and chlordanes, chlordanes and Metals 1, and PCBs and Metals 1, which showed insignificant correlations (Table 19).
Liver and bile chemistry--Multiple variables related to contaminant exposure in bile and liver tissue were also identified as risk factors of hepatic disease; these measurements document actual exposure to and bioaccumulation of contaminants (Table 17). At least one biliary FAC measurement (FACs-H or FACs-L) was a risk factor for all detected hepatic lesions in English sole, with the exception of necrosis. These findings parallel the significance of the risk factors for AH exposure (LAHs and HAHs) in sediment and stomach contents described above. The proportion of the intersite variation in lesion prevalence accounted for by biliary FACs ranged from 13 to 35%, depending upon the lesion type. Both DDTs and PCBs in liver tissue were also risk factors for neoplasms, foci of cellular alteration, SDN, and proliferative lesions, explaining between 24 and 66% of the prevalence variation among these lesion categories. Liver DDTs was also a risk factor for necrosis. In contrast to the risk factors identified in sediment, neither chlordanes, dieldrin, nor Metals 1 in liver tissue were risk factors. Again, it is also important to note here that some of the identified risk factors in bile and liver tissue were correlated with one another (Table 20). Specifically, levels of FACs-L were significantly correlated with both liver DDTs and PCBs, while DDTs and PCB levels in liver also showed a significant correlation.
Summary--Logistic regression analyses in English sole relating hepatic lesion prevalences to risk factors of contaminant exposure showed a number of risk factors in the various compartments measured (sediments, stomach contents, bile and liver tissue); the intercorrelations among these risk factors within each compartment complicate interpretation and prohibit quantification of their relative significance. When consistency is expected, risk factors that consistently showed significance to hepatic disease among the various compartments should be regarded as etiologic factors of greater significance. For example, the LAHs and HAHs were consistently identified as risk factors for most of the hepatic lesions, whether measured in sediments, stomach contents, or bile (as reflected in biliary FACs-H or FACs-L). In fact, the correlations among the compartments for these risk factors were generally strong and significant, except for a lack of correlation between AHs in stomach contents and FACs in bile (Table 21). Strong correlations among the compartments were also demonstrated for total PCBs, which was a consistent risk factor for neoplasms, foci of cellular alteration, SDN, and proliferative lesions, based on values in the sediment and liver tissue, and for foci of cellular alteration, SDN, and proliferative lesions based on concentrations in stomach contents. The DDTs were a less consistent and less frequently appearing risk factor than AHs or PCBs, being significant only for SDN and proliferative lesions among all three compartments in which DDTs were measured. However, this lower consistency is reflected in the lower and insignificant correlation coefficients shown for DDT levels among the liver tissue and stomach content compartments (Table 21); therefore, fully consistent associations among the compartments between DDT levels and lesion prevalences should not be expected. The DDTs were significantly correlated most commonly with only the PCBs, and not the AHs in the separate compartments. Of the two pesticides identified as risk factors (chlordanes and dieldrin), only chlordanes showed any consistency among the compartments, being identified with both SDN and proliferative lesions, but only among sediments and stomach contents, and not for liver tissue. Dieldrin in sediments was a risk factor for neoplasms and proliferative lesions, but it was not reinforced as a risk factor indicating bioaccumulation (liver levels). In fact, the correlation between dieldrin levels in sediment and liver was negative. In parallel with this finding is the inconsistent pattern of association between levels of Metals 1 and hepatic lesions among the compartments. While
Metals 1 in sediments was a risk factor for all detected liver lesions, there was little consistency among the compartments measured; in fact, the correlation coefficients for levels of Metals 1 in sediment and liver, and Metals 1 in stomach contents and liver were also negative (Table 21). Moreover, levels for Metals 1 were highly correlated (in sediments) with other significant risk factors that were consistently identified among all compartments: LAHs, HAHs, and PCBs (Table 18).
Starry flounder--Identical statistical treatment of the data for starry flounder showed a number of contaminant exposure-related risk factors for the hepatic lesions foci of cellular alteration, SDN, and especially hydropic vacuolation (Tables 22-24).
Sediment chemistry--Among chemicals measured in sediments, the LAHs, HAHs, PCBs and Metals 1 were highly significant risk factors for both SDN and hydropic vacuolation, accounting for between 12 and 32% of the intersite variation in prevalence of these lesions, respectively (Table 22). For hydropic vacuolation, hexachlorobenzene was also a risk factor. For the lesion SDN, sediment LAHs, HAHs, and PCBs explained approximately equal proportions of the variation in prevalence. Levels of these chemicals in sediments were also moderately to strongly intercorrelated (Table 25). Metals 1 was also a risk factor for SDN; however, Metals 1 levels in sediment were strongly intercorrelated with LAHs, HAHs, PCBs, and hexachlorobenzene. Approximately equal proportions of the variation in prevalence of hydropic vacuolation were accounted for by LAHs, HAHs, PCBs, hexachlorobenzene, and Metals 1; these findings were consistent with the significant intercorrelations among all these risk factors. The relationships among the risk factors for SDN and hydropic vacuolation (Table 25) suggest that the LAHS, HAHs, PCBs, and Metals 1 form a distinct complex of risk factors. For hydropic vacuolation, hexachlorobenzene is a risk factor most strongly correlated with PCBs and Metals 1 levels.
Stomach contents chemistry--As indices of dietary exposure, several chemical groups measured in stomach contents were identified as risk factors of hepatic disease (Table 23), with somewhat different results from those obtained for risk factors in sediments or liver tissue. For example, no risk factors in stomach contents were identified for the lesion category SDN, whereas several risk factors were shown for the lesion categories foci of cellular alteration and hydropic vacuolation. The highly intercorrelated chemicals, DDTs and PCBs (Table 26), were risk factors for foci of cellular alteration, and explained a large proportion of the intersite prevalence variation. Dietary exposure to HAHs and total AHs was a risk factor for hydropic vacuolation, as was DDTs, PCBs, and chlordanes. Levels of these chlorinated hydrocarbon groups were highly and significantly intercorrelated (Table 26). However, none of these chlorinated compounds were correlated with any measure of AH exposure in stomach contents.
Liver and bile chemistry--Several risk factors of hepatic disease reflecting actual chemical uptake and exposure were identified in bile and liver tissue (Table 24) and were generally the same as those demonstrated in sediment; risk factors were identified only for the lesions SDN and hydropic vacuolation, with the single exception that Metals 1 was a risk factor for necrosis. For the lesion SDN, AH exposure (as reflected by levels of biliary FACs) was not a risk factor, while the highly intercorrelated (Table 27) measures of PCBs, chlordanes, and dieldrin were (9-20% of prevalence variation explained). In contrast to the results for sediment, Metals 1 in liver tissue was not associated with SDN in starry flounder, as reflected by its lack of correlation with any of the chlorinated compounds or AHs (Table 27).
For hydropic vacuolation, AH exposure was confirmed as an important risk factor (Table 24), as reflected by the correlated measures (Table 27) of mean levels of biliary FACs-H and FACs-L. Other identified risk factors in liver tissue for this lesion type, accounting for large proportions of the prevalence variation (25-36%) were the chlorinated hydrocarbons DDTs, PCBs, chlordanes and dieldrin, all of which were highly intercorrelated (Table 27).
Summary--Multiple risk factors were identified for hepatic lesion prevalences in starry flounder in the compartments measured. This was especially true for the most commonly detected lesion, hydropic vacuolation, where exposure to AHs and various covarying chlorinated hydrocarbons (especially DDTs and PCBs in the separate compartments) was, in general, consistently associated with increased prevalence. Chlordanes, hexachlorobenzene, and dieldrin were identified as risk factors in only one or two compartments. However, consistency among the compartments for these risk factors would not be expected, based on the generally insignificant level of intercompartmental correlation among these variables (Table 28). The fact that DDTs only in liver and stomach contents were risk factors for hydropic vacuolation is consistent with the lack of correlation between DDT levels in these compartments to those in sediment (Table 28). Similarly for Metals 1, there was no correlation among any of the compartments measured, although Metals 1 levels strongly covaried with other chemical risk factors for hydropic vacuolation identified in sediments. For the lesion type SDN, only PCBs were risk factors in at least two compartments (sediments and liver), consistent with the generally insignificant level of correlation among the compartments for the other risk factors identified, such as the AHs, Metals 1, chlordanes and dieldrin. The intercorrelated measures of PCBs and DDTs in stomach contents were the only risk factors for foci of cellular alteration, reflecting the rare occurrence of this lesion category in starry flounder.
Hornyhead turbot--The only chemical exposure-related risk factor positively associated with increased risk of any hepatic lesion in hornyhead turbot was dieldrin in sediments (n = 13 for sediment chemistry except for metals measures, where n = 11; for stomach contents chemistry, n = 6, except for metals measures where n = 3; for liver chemistry, n = 12 except for metals measures, where n = 9; for FACs in bile, n = 11). This factor was associated with increased risk of foci of cellular alteration, and accounted for 45% (p = 0.047) of the intersite variation in prevalence (other data not shown). Mean biliary FACs-H or FACs-L were not risk factors for any hepatic lesion category.
White croaker--Results of logistic regression analyses for hepatic disease risk due to contaminant exposure as determined on a by-site basis in white croaker are shown in Tables 29-31. Risk factors were identified for all lesion types except neoplasms.
Sediment chemistry--At least one chemical class or group measured in sediment was associated with increased risk of occurrence for all liver lesion types, except neoplasms (Table 29). For foci of cellular alteration, sediment LAHs and HAHs explained 23-28% of the intersite variation in lesion prevalence. In the case of SDN, the chlorinated hydrocarbons represented by DDTs, PCBs, and chlordanes were risk factors accounting for between 8 and 31% of the prevalence variation, with Metals 1 independently accounting for 26%. The only risk factor for hydropic vacuolation was hexachlorobenzene, while DDTs and PCBs significantly influenced the prevalence of the proliferative lesions. Risk factors for necrosis were both aromatic hydrocarbon measures (17-21% of variation explained) and PCBs (16%). The AH measures in sediment were highly intercorrelated, as were Metals 1 with HAHs and total AHs (Table 32). The AH measures were also significantly correlated with PCBs and hexachlorobenzene. PCBs were also strongly correlated with Metals 1 in sediments. Chlordanes correlated most strongly with Metals 1 and PCBs, the latter of which in turn was correlated with DDTs. These complex interrelationships are reflected in the patterns of co-occurrence of the significant risk factors in sediments, such that the AHs formed a group of risk factors fairly distinct from the chlorinated hydrocarbons (excepting PCBs, with which they were weakly but significantly correlated); DDTs and PCBs only occurred together as risk factors; chlordanes, Metals 1, and PCBs were co-occurring risk factors for SDN as well as being significantly intercorrelated; and hexachlorobenzene occurred independent of other risk factors because it was not highly correlated with any other chemical parameter measured in sediment, showing only weak, yet significant correlations with the AHs (Table 32). Overall, AH levels in sediments were linked to foci of cellular alteration and necrosis, whereas the groups within the chlorinated hydrocarbons, and Metals 1 were risk factors for SDN, hydropic vacuolation, proliferative lesions, and necrosis.
Stomach contents chemistry--Risk factors identified in stomach contents were mainly associated with SDN and proliferative lesions, and were primarily represented by classes within the chlorinated hydrocarbons, accounting for between 8 and 39% of the intersite variation in lesion prevalence (Table 30). Aromatic hydrocarbons were not associated with increased risk for any lesion category. Risk factors accounting for substantial proportions of the variation in prevalence for SDN were DDTs, PCBs, chlordanes, and Metals 1. Co-occurrence of the chlorinated hydrocarbons as risk factors for SDN reflected their high degree of intercorrelation (Table 33). The emergence of Metals 1 as a risk factor for SDN was anomalous, considering its lack of correlation with any of the chlorinated hydrocarbons in stomach contents that were also risk factors for this lesion (Table 33), as well as the absence of any correlation between liver and stomach levels for Metals 1 (Table 34). Contrasting with the results for sediments and liver tissue, hexachlorobenzene in stomach contents was not a risk factor for hydropic vacuolation, reflecting the lack of correlation between its levels in sediment and stomach contents (Table 34). The correlated measures of DDTs and PCBs (Table 33) were risk factors for proliferative lesions, while hexachlorobenzene appeared to be an independent risk factor, since it was not correlated with DDTs or PCBs. Increased risk of liver necrosis was associated only with hexachlorobenzene in stomach contents.
Liver and bile chemistry--Several chemicals measured in bile and liver tissue were also risk factors of hepatic disease (Table 31). Levels of biliary FACs-H or FACs-L were associated with foci of cellular alteration, SDN, hydropic vacuolation, proliferative lesions and necrosis, accounting for between 4 and 32% of the intersite variation in lesion prevalences. These findings agree with the results of logistic regression analyses of the risk of foci of cellular alteration and SDN occurrence in individual fish with respect to levels of biliary FACs. However, FACs in bile were not significant risk factors for hydropic vacuolation, proliferative lesions or necrosis in the analysis done on individual fish. Chemical risk factors in liver tissue included DDTs (SDN, proliferative lesions and necrosis), PCBs (SDN and hydropic vacuolation), hexachlorobenzene (hydropic vacuolation), Metals 1 (SDN and proliferative lesions) and Metals 2 (SDN). The presence of DDTs as a risk factor for SDN and proliferative lesions paralleled the results for sediments, and reflected the strong correlation between levels of DDTs in liver and sediment (Table 34). The lack of accord between the lesion types associated with PCB levels in liver tissue and sediments may be partially accounted for by the relatively low, yet significant correlation between PCBs levels in sediments and liver tissue in white croaker (Table 34). Hexachlorobenzene in liver tissue was a risk factor only for hydropic vacuolation, and this variable was not correlated with any other chemical measure in liver tissue. Metals 1 and Metals 2 were risk factors for both SDN and proliferative lesions, but also strongly covaried with DDTs (Table 35), a highly significant risk factor for both of these lesion types.
Summary--A number of risk factors were consistently associated with hepatic lesion types in white croaker, except for neoplasms. For foci of cellular alteration, AHs were identified as risk factors both in sediments (LAHs and HAHs) and bile (FACs-L). The absence of stomach content AHs as risk factors for foci of cellular alteration can be attributed to the general lack of correlation among the AHs in sediments, stomach contents, and as reflected in biliary FACs (Table 34), in addition to the far lower sample size (18) represented for stomach contents than that for sediments (43) and biliary FACs (44).
Consistently identified risk factors for SDN in all compartments were the highly correlated risk factors: DDTs and PCBs (Table 34). Metals 1 was also a consistent risk factor for this lesion, possibly due to its correlation with PCBs and chlordanes in sediments and DDTs in liver tissue. Chlordane was identified less consistently as a risk factor for SDN, and accounted for a lower proportion of the prevalence variation, even though levels among the compartments were highly correlated (Table 34). Biliary FACs-L and FACs-H were the only measures of AH exposure that were risk factors for SDN; this is consistent with the lack of significant correlation between biliary FACs and sediment and stomach content AHs in this species (Table 34).
For hydropic vacuolation, only hexachlorobenzene was a consistently identified risk factor (sediments and liver tissue), with the only AH exposure-related risk factor being biliary FACs-L; PCBs were a risk factor for hydropic vacuolation, but only in liver tissue.
As was true for SDN, and reflecting their significant intercompartmental correlations (Table 34), the covarying risk factors of DDTs and PCBs were most consistently shown as risk factors for nonneoplastic proliferative lesions; but biliary FACs-H and FACs-L were the only measures of AH exposure shown to be risk factors for proliferative lesions. Hexachlorobenzene (stomach contents), and Metals 1 (liver) were identified only once as risk factors for this lesion.
For necrosis, the risk factors explaining the highest proportion of intersite variation in prevalence were the sediment AHs, a relationship that was at least partially supported by the existence of biliary FACs-L as a risk factor. No other risk factors for this lesion type were consistently identified in the sediments, bile/liver tissue or stomach contents; PCBs in sediments (correlated with sediment AHs), hexachlorobenzene in stomach contents, and DDTs in liver were identified once as risk factors for this lesion type.
Black croaker--This species was not selected as a primary target species in the NBSP until 1987. Therefore, there were only a few sites sampled in 1987 and 1988 that were analyzed for sediment (7), liver tissue (4), stomach contents chemistry (0), and biliary FACs (4). Consequently, a logistic regression analysis of hepatic disease risk relative to site-of-capture risk factors was not considered valid and was not done.
Assessment of Variables Related to Contaminant Exposure (Determined on a By-Site Basis) as Risk Factors for Renal Lesions
Flathead sole--Results of logistic regression analyses for renal disease risk due to contaminant exposure as determined on a by-site basis in flathead sole are shown in Tables 36 and 37. Risk factors were identified for all kidney lesion categories.
Sediment chemistry--Several chemicals measured in sediments were indicated as risk factors for all three renal lesion categories (Table 36). Both dieldrin and Metals 2 levels were associated with prevalences of the proliferative lesions; however, since dieldrin was detected at a very low level only at one site in a single year (Commencement Bay), it is not considered a risk factor (see Discussion). For the necrotic lesions, LAHs and HAHs explained 39-42% of the prevalence variation, with DDTs, PCBs, and Metals 1 accounting for an even higher proportion (61, 57, and 45%, respectively). Both measures of sediment AHs, in addition to PCBs, were also risk factors for the sclerotic kidney lesions; however, the LAHs, HAHs, DDTs, and PCBs were strongly intercorrelated (Table 38). Metals 1 may represent a risk factor for necrotic lesions that is not independent of the AHs and PCBs, due to its significant correlation with them. However, considering its lack of correlation with the AHs or CHs, Metals 2 is clearly a risk factor independent of these chemical groups (Table 38).
Stomach contents chemistry--Risk factors in stomach contents (n = 7) were restricted to DDTs, accounting for 67% (p = 0.006) of the variation in prevalence of sclerotic lesions.
Liver and bile chemistry--Several risk factors were demonstrated in the bile and liver tissue (Table 37). Consistent with the relationship demonstrated with sediment AHs, biliary FACs-H was a risk factor for both necrotic and sclerotic lesions. For the proliferative lesions, risk factors measured in liver tissue were DDTs, PCBs, dieldrin and chlordanes. The same chlorinated organics, in addition to hexachlorobenzene, were also risk factors for the sclerotic kidney lesions. However, most of these chlorinated organic compounds were strongly intercorrelated in liver (Table 39) and were not independent risk factors; however, hexachlorobenzene in liver appeared to be independent, as it did not correlate significantly with levels of the other chlorinated contaminants. Levels of biliary FACs-H also strongly covaried with all of the chlorinated hydrocarbons that were significant risk factors, with the exception of hexachlorobenzene.
Summary--Although several indices of contaminant exposure were risk factors for all three renal lesion categories in flathead sole, few of these were consistently identified among the compartments measured. For example, no risk factors for any of the lesion categories were consistently identified among the sediment, liver tissue, and stomach contents compartments; however, this may be partially explained by the low sample size (n = 7) for the stomach contents. Only the PCBs showed any strong degree of correlation among the three measured compartments; correlations of levels of other specific chemicals among the compartments were quite low, and in many cases, were negative (Table 40). Of the risk factors showing strong correlations in at least two of the compartments, and hence where consistency might reasonably be expected, exposure to and uptake of HAHs was a risk factor for both the necrotic and sclerotic lesions. On the same basis, PCB exposure was a consistent risk factor for sclerotic lesions. Although DDTs in sediment was not a risk factor for the sclerotic lesions, it was in compartments indicating actual exposure and uptake (stomach contents) and bioaccumulation (liver). For the proliferative lesions, no risk factor was identified in more than one compartment; however, the intercorrelated measures of PCBs, DDTs and chlordanes in liver (indicating bioaccumulation) were risk factors for this lesion category. Finally, although necrotic lesions were strongly associated with Metals 1 in sediments, but not with these metals in liver or stomach contents, the intercompartmental correlations for the metals groups were not significant (Table 40).
Sediment chemistry--Of the chemicals measured in sediment (Table 41), both classes of AHs, as well as PCBs and Metals 1 were risk factors for the proliferative lesions, explaining 10-19% of the intersite variation in prevalence. No risk factors were associated with the necrotic or sclerotic lesions. As previously described for the liver lesions, these relationships between renal lesion prevalences and sediment levels of these chemical groups must be considered in light of the very strong correlations that exist among the LAHs, HAHs, PCBs and Metals 1 (Table 18).
Stomach contents chemistry--The only risk factor identified in stomach contents (n = 9); was chlordanes, which explained 65% of the variation (p = 0.009) among site-specific prevalences for the necrotic renal lesions.
Liver and bile chemistry--Chemicals measured in liver tissue/bile were also risk factors for the renal lesion categories in English sole (Table 42). Biliary FACs-H, PCBs, dieldrin, and
Metals 1 in liver were risk factors for the proliferative lesions. The highly intercorrelated measures (0.978 < R < 0.995, p = 0.0001) of liver hexachlorobenzene and chlordanes were also positive risk factors for both the necrotic and sclerotic lesions, while dieldrin explained a significant proportion of the prevalence variation for the sclerotic lesions.
Summary--Several chemical exposure risk factors for the kidney lesion categories were consistently identified among the two compartments that contained a reasonable number of samples (sediments and liver/bile). Using this criterion for consistency, exposure to AHs, PCBs, or
Metals 1 explained a large proportion of the prevalence variation for the proliferative lesions. Chlordanes were a consistent risk factor (stomach contents and liver) for the necrotic lesions. No other risk factors for other renal lesion categories were consistently identified, although the highly intercorrelated (R = 0.871, p = 0.0001, n = 13) measures of hexachlorobenzene and chlordanes were strong risk factors representing hepatic bioaccumulation for both the necrotic and sclerotic lesions. Hepatic bioaccumulation of dieldrin was also a risk factor for the sclerotic kidney lesions.
Sediment chemistry--The only chemical measure in sediments that showed a significant contribution to the explanation of the variation in prevalence of any kidney lesion was Metals 1; this variable accounted for 31% (p = 0.001) of the prevalence variation for the necrotic lesions (33 sites were analyzed for AHs and CHs; 19 were analyzed for Metals 1 and Metals 2).
Stomach contents chemistry--The single risk factor associated with renal lesions was chlordanes, which explained 15% (p = 0.009) of the intersite prevalence variation for the necrotic lesions (number of sites = 15). Unlike the relationship shown for the sediment compartment, Metals 1 in stomach contents was not a risk factor for the necrotic lesions. This relationship is consistent with the negative correlation between levels of these metals in sediment and stomach contents (Table 28).
Liver and bile chemistry--Chemical risk factors in liver tissue were the highly intercorrelated measures (Table 27) of DDTs and PCBs, accounting for 4-14% of the prevalence variation for both the necrotic and sclerotic lesions (Table 43). Dieldrin, which covaried with DDTs and PCBs (Table 27), also was a risk factor for the necrotic lesions. The absence of Metals 1 in liver tissue as a risk factor for the necrotic lesions was expected, considering the negative correlation between sediment and liver tissue levels for this group of trace metals (Table 28).
Summary--Necrotic lesions were associated with potential exposure to compounds in the Metals 1 group (sediments), with uptake of chlordanes (stomach contents), and with bioaccumulation of PCBs, DDTs, and dieldrin (liver). Sclerotic lesions were only associated with bioaccumulation of DDTs and PCBs; no risk factors for proliferative lesions were identified. Although there was little consistency among the compartments assessed for the few risk factors identified in the individual compartments (sediment Metals 1, liver DDTs, PCBs and dieldrin, and chlordanes in stomach contents), the generally insignificant intercompartmental correlations for these chemicals (Table 28) indicate that consistency would not be expected. Only PCBs in liver vs. sediments, and in liver vs. stomach contents showed a strong degree of correlation. The strong correlations among the PCBs, DDTs and dieldrin in liver suggest that these risk factors of renal disease are not independent.
Hornyhead turbot--No risk factors for renal disease in hornyhead turbot were identified among the chemicals measured in sediments, liver tissue, bile, or stomach contents (n = 13 for sediment chemistry except for metals measures where n = 11; for stomach contents chemistry, n = 6, except for metals measures where n = 3; for liver chemistry, n = 12 except for metals measures, where n = 9; for FACs in bile, n = 11).
Sediment chemistry--Of the contaminants measured in sediment, only chlordanes and DDTs were risk factors, accounting for 6% (p = 0.044) of the intersite variation in prevalence for both the necrotic and sclerotic kidney lesions in the case of chlordanes and 5% (p = 0.017) of the intersite variation for the proliferative lesions in the case of DDTs (n = 43 for measures of AHs and CHs, n = 23 for metals measures). These sediment chemicals were not correlated with one another (
Table 32).Stomach contents chemistry--No indices of contaminant exposure measured in stomach contents were associated with any renal lesion category (n = 18 for measures of AHs and CHs, n = 8 for metals measures).
Liver and bile chemistry--In liver, DDTs were a risk factor for the necrotic and proliferative lesions, accounting for 9% (p = 0.012) and 8% (p = 0.006) of the variation in prevalences, respectively (n = 31 for measures of AHs and CHs, n = 13 for metals measures). The only chemical group measured in liver tissue that was associated with the sclerotic lesions was Metals 1, explaining 16% (p = 0.028) of the intersite prevalence variation. Neither measure of biliary FACs was associated with any kidney lesion (n = 44).
Summary --The only risk factor for renal disease consistently identified in at least two of the measured compartments was DDTs (liver and sediments), which was associated with prevalences of the proliferative lesions. These DDT measures were highly intercorrelated (Table 34). For the sclerotic lesions, chlordanes would be expected to exist as a consistent risk factor among the compartments based on the significant intercompartmental correlation; however, it was only identified as a risk factor in sediments. Similarly, chlordanes was an inconsistent risk factor for necrosis. Metals 1 as a risk factor for sclerosis must be seriously regarded since it was identified in a compartment indicating bioaccumulation (liver); the same applies to DDT as a risk factor for necrosis.
Black croaker--Because of the low number of sampling sites and low prevalences detected, a logistic regression analysis of kidney disease risk relative to chemical exposure and biological parameters determined on a site basis was not performed for this species.
Temporal trends in hepatic lesion prevalence were tested in the following major target species at sites that were annually sampled at least four times between 1984 and 1988: English sole (Bodega Bay), starry flounder (San Pablo Bay, Southampton Shoal, Hunters Point, Bodega Bay, Coos Bay), hornyhead turbot (Dana Point), barred sandbass (south San Diego Bay) and white croaker (Bodega Bay, Hunters Point, Southampton Shoal, Dana Point, San Pedro outer harbor). All idiopathic liver lesion categories diagnosed in the respective species at the indicated sites of capture were tested for temporal trends in lesion prevalences by the Spearman rank correlation method (Table 44). No significant monotonic increases or decreases in any hepatic lesion category in any of the tested species and sites were detected. Logistic regression analyses of the same data, while adjusting for mean age and gender ratio at each of the sites, also failed to show any significant temporal trends (i.e., increase or decrease) in prevalences for any hepatic lesion category (data not shown).